Recently, millimeter-wave (mmWave) 5G localization has been shown to be to provide centimeter-level accuracy, lending itself to many location-aware applications, e.g., connected autonomous vehicles (CAVs). One assumption usually made in the investigation of localization methods is that the user equipment (UE), i.e., a CAV, and the base station (BS) are {time} synchronized. In this paper, we remove this assumption and investigate two two-way localization protocols: (i) a round-trip localization protocol (RLP), whereby the BS and UE exchange signals in two rounds of transmission and then localization is achieved using the signal received in the second round; (ii) a collaborative localization protocol (CLP), whereby localization is achieved using the signals received in the two rounds. We derive the position and orientation error bounds applying beamforming at both ends and compare them to the traditional one-way localization. Our results show that mmWave localization is mainly limited by the angular rather than the temporal estimation and that CLP significantly outperforms RLP. Our simulations also show that it is more beneficial to have more antennas at the BS than at the UE.

中文翻译：

---

5G 毫米波系统的单锚双向定位边界

最近，毫米波 (mmWave) 5G 定位已被证明可提供厘米级精度，适用于许多位置感知应用，例如联网自动驾驶汽车 (CAV)。在定位方法的研究中通常做出的一个假设是用户设备(UE)，即CAV，和基站(BS){时间}同步。在本文中，我们去除了这个假设并研究了两种双向定位协议：(i) 往返定位协议 (RLP)，其中 BS 和 UE 在两轮传输中交换信号，然后使用信号实现定位第二轮收到；(ii) 协作定位协议 (CLP)，其中使用两轮接收到的信号实现定位。我们推导出在两端应用波束成形的位置和方向误差界限，并将它们与传统的单向定位进行比较。我们的结果表明，毫米波定位主要受角度而不是时间估计的限制，并且 CLP 明显优于 RLP。我们的模拟还表明，在 BS 处拥有更多天线比在 UE 处拥有更多天线更有利。